OpenAI and SpaceXAI unveiled new artificial intelligence models on Wednesday, with OpenAI introducing a voice system that can listen and speak at the same time and SpaceXAI launching a coding-focused model built for agentic software tasks.
OpenAI’s GPT-Live is designed to make voice interaction feel less like a turn-based chatbot exchange and more like a real conversation, while SpaceXAI’s Grok 4.5 is aimed at developers and enterprises seeking faster, cheaper and more autonomous support for coding, technical workflows and task execution.
OpenAI pushes voice AI into real-time conversation
OpenAI said GPT-Live is built on a full-duplex architecture, allowing the model to listen and speak at the same time instead of waiting for a user to finish before responding.
The company said the design makes ChatGPT Voice more fluid and natural, enabling the assistant to give brief signals that it is following along, respond in quick exchanges or remain silent while a user pauses to think.
The shift is meant to reduce the awkward delays and interruptions that have limited earlier voice systems, where silence detection often shaped the pace of the conversation.
OpenAI is rolling out two versions of the system, with GPT-Live-1 powering the new Voice experience for paid users and GPT-Live-1 mini serving Free users as a lighter version aimed at broader access.
For more complex requests, the system can route work to frontier models in the background, allowing the voice experience to continue while deeper reasoning, search or advanced tasks are handled separately.
The launch extends OpenAI’s push into voice-based AI, building on its earlier audio models for developers and moving ChatGPT closer to live use cases ranging from everyday conversation to workplace tasks.
SpaceXAI targets developers with Grok 4.5
SpaceXAI launched Grok 4.5 as its latest push into developer tools and autonomous AI workflows, expanding its effort to gain ground in the fast-growing market for AI coding tools.
The company said Grok 4.5 is its most intelligent model to date, built for coding, agentic tasks and technical workflows. The model was trained across tens of thousands of Nvidia GB300 graphics processing units, with an emphasis on data filtering, deduplication and quality scoring.
Grok 4.5 is available through Grok Build, Cursor and the SpaceXAI console, with European access expected in mid-July, and is priced at $2 per million input tokens and $6 per million output tokens. Elon Musk described it as an Opus-class model that is faster, more token-efficient and lower cost.
AI race turns toward everyday use
Together, the announcements point to a new phase in the AI race, where companies are moving beyond advanced reasoning chatbots and focusing more heavily on systems that can work inside real-world routines.
OpenAI is trying to make voice interaction feel more natural and immediate, while SpaceXAI is targeting developers and businesses that need AI to support coding, manage technical tasks and operate across software workflows.
The broader shift is toward models that do more than generate answers, with companies racing to build AI systems that can communicate, assist and act across the tools people already use.
